| Popis: |
The occurrence of Appalachian cold air damming (CAD) is often associated with significant sensible weather impacts throughout the Carolinas and Virginia. CAD sensible weather, defined as below-normal maximum temperatures, overcast skies, fog, and reduced visibility, can often persist in the damming region for days. Furthermore, the confinement of a dome of low-level cold air along the eastern slopes of the Appalachians can create an environment ideal for freezing rain and sleet. Such potentially hazardous conditions necessitate accurate and timely forecasting in order to properly warn the public. Despite this need for reliable CAD prediction, accurate forecasting of the magnitude of sensible weather impacts and the timing of CAD demise are extremely challenging. Furthermore, many operational numerical weather prediction models do not accurately simulate CAD events. One of the more common problems associated with model forecasts is the premature erosion of the CAD cold dome and the underestimation of the duration of CAD sensible weather. A better understanding of the physical processes involved in cold dome erosion would lead to improved erosion forecasting by increasing forecaster awareness of the signs of CAD erosion and by isolating physical processes that may be poorly handled by operational models. One objective of this research is to describe the common synoptic patterns evident during the erosion period of CAD events. These synoptic erosion scenarios are representative of the physical mechanisms contributing to cold dome erosion. To accomplish this objective, 89 classical CAD events detected by the algorithm of Bailey et al. (2003) were grouped according to similarities in sea-level pressure and surface potential temperature fields. Composite maps of each group were created, resulting in five CAD erosion scenarios: (1) Northwestern Low, (2) Cold Frontal Passage, (3) Coastal Low, (4) Residual Cold Pool, and (5) Southwestern Low. High levels of statistical significance were associated with the dominant synoptic features in all scenarios except the Southwestern Low, suggesting that the remaining four scenarios effectively represented distinct erosion patterns. Several potential erosion mechanisms could be inferred from the evolution of dominant synoptic features in each erosion scenario. Multiple erosion mechanisms may be simultaneously contributing to cold dome erosion for any one given scenario. The second objective involved detailed case studies of three CAD events in order to more closely examine erosion mechanisms. The first case was an example of the Coastal Low erosion scenario, the second case was representative of the Northwestern Low erosion scenario, and erosion of the third case was multi-faceted and not clearly classifiable. Detailed examination of observations and EDAS analyses revealed that multiple erosion mechanisms were contributing to the weakening of the capping inversion above the cold dome and promoting the erosion of the CAD event. In the Coastal Low case, cold advection aloft led to a decrease in the potential temperature difference across the inversion, indicating erosion from the top of the cold dome down to the surface. Comparatively, the inland progression of a coastal front, in association with surface divergence, corresponded to an increase in surface temperatures and a weakening of the inversion during the Northwestern Low case study. This development was more indicative of erosion from the surface upwards. The performance of NCEP Eta Model forecasts during the three CAD events was evaluated. Consistent with the findings of previous research, the Eta Model eroded the cold dome prematurely for all three cases, and surface temperature were consistently overestimated. Control run simulations using the PSU/NCAR MM5 Model were performed for the Coastal Low and Northwestern Low cases. The control run simulations showed improved accuracy over the Eta Model forecasts in the representation of CAD erosion. However, erosion was still premature in comparison to observations. For both CAD events, overestimated values of shortwave radiation appeared to correlate with the decrease in model inversion strength. Finally, two sets of sensitivity tests for the Coastal Low case using the MM5 model were designed to test the sensitivity of model performance to alterations in certain physics parameterizations. The first sensitivity test involved altering the model values of cloud albedo, based on speculations that the overestimation of surface temperatures in the Eta Model was a result of the interaction between clouds and shortwave radiation. As hypothesized, the simulation in which cloud albedo was decreased produced the warmest surface temperatures. The second sensitivity test involved altering the PBL schemes used in the MM5 control run. It was found that the simulated vertical structure of the atmosphere did vary according to the PBL scheme, as anticipated. Although some of the alterations made to the MM5 control run for the sensitivity tests did improve the representation of certain features, the model continued to erode the cold dome earlier than observed. Comparisons of the MM5 simulations to the corresponding Eta Model forecasts revealed that none of the alterations to the MM5 control run produced the magnitude of error evident in the Eta Model. |
